AI curriculum to be introduced in the Medicine Programme

The project “Integrating AI in medical education”, funded through the Marianne and Marcus Wallenberg Foundation’s educational development initiative (WASP-ED), was initiated by Marcus Krantz, Researcher in Systems Biology at Örebro University. He soon contacted Johan Sundström, Professor of Epidemiology at Uppsala University.

“I’ve worked quite a bit with AI in healthcare, which is why Marcus got in touch when the project began to involve several universities. Together with colleagues from both universities, we have developed a curriculum on AI that can be implemented in all medical programmes across the country,” says Johan Sundström.

It has now been confirmed that this will indeed happen. Programme directors from all seven medical programmes in Sweden have agreed to proceed with the project in three phases, with the introduction of the curriculum planned for the autumn semester of 2026.

“It’s very exciting that we’ve reached this point. AI is an area where development is moving incredibly fast. The decision to implement the curriculum over the coming year is an important step towards accelerating the work to ensure a relevant and up-to-date syllabus.”

Bringing the content to life

The proposed curriculum contains sixteen digital modules divided into four levels: introduction, theory, application and specialisation.

“All material will be digital and designed so that each university can implement the modules in any order. We know that bringing the different elements to life is crucial for learning. The working group now continuing with the modules will therefore provide the theoretical framework, while it will be up to each institution to introduce practical components such as lab exercises, group work or similar,” says Johan Sundström.

Although part of the work on the curriculum has already been completed and will continue up to the implementation, it will also require continuous updating, Johan explains.

“As mentioned, development is incredibly rapid. We’ve structured the course so that the first half focuses on the technology itself – how various AI models work, how they are trained and what they can and cannot do. That’s the area evolving most quickly. The second half will deal more with clinical, ethical and legal issues – for instance, which patient groups the models have been trained on and whether there are biases at risk of being reproduced. Such questions are more timeless, so to speak.”

Enormous potential – but also challenges

Why is a curriculum on AI important for physicians and potentially other healthcare professionals? To illustrate this, Johan Sundström draws on his own experience as a cardiologist.

“If we take my field, ECG is a clear example of where AI is used. There are currently two ECG algorithms used globally. They are hard-coded, meaning they are programmed according to fixed rules – the same rules I use as a physician when interpreting the curves,” he explains.

“But the AI models now being developed can make diagnoses based on patterns that humans don’t even understand. They find information in ECG signals that we simply cannot see, and that means we must relate to the algorithms in a new way. We need to better understand what the model has been trained on and assess its reliability. This means that we place higher demands on AI algorithms than on the hard-coded ones that we are used to today.”

Interdisciplinary collaboration

At present, this AI education will mainly be offered within the country’s medical programmes. However, Johan Sundström also sees potential benefits for other healthcare professions.

“Absolutely. In essence, the questions are the same for nurses or physiotherapists, for example. Another consideration is whether the course should also be offered to practising physicians. That would be possible, but it might require a more independent web-based version without the local variations that will be shaped by lab sessions and similar activities.”

The next step will be to further develop the course modules so that implementation can begin in autumn 2026. This work will also involve contributors from fields beyond healthcare.

“I was a bit worried that just a few of us would have to record all the lectures for the various modules ourselves, but when I reached out to colleagues from other faculties, I received great responses. Engineers, lawyers, philosophers and others have all shown interest in contributing. This gives us greater breadth and quality overall. Those of us from the healthcare field can contribute with clinical understanding, while others bring technical, legal and ethical perspectives, for instance,” says Johan Sundström.

Robin Widing